Machine reasoning
Powerful and intelligent machine reasoning technologies are a critical enabler of today’s intent-based networks. In an increasingly complex world of 5G network operations, machine reasoning technologies apply human-like common sense to analyze and translate vast knowledge and learned network data into clear explainable insights. For service providers, this is key to ensuring that network operations continuously align with business intent, meet increased customer expectations and deliver on new 5G revenue opportunities.
What is machine reasoning?
Machine reasoning is a subset of artificial intelligence (AI) which enables machines to act with agility and relevancy according to learned data. Similar to human intelligence, artificial intelligence is not only defined by the ability to learn, but it must also have the ability to reason – to bring relevance to decisions and actions, and to adapt in different situations.
Therefore, while machine learning is typically applied to learn complex functions using vast amounts of data, machine reasoning models are applied to integrate intent within those processes.
The difference between machine learning and machine reasoning
While machine learning is typically applied to learn complex functions using vast amounts of data, such as learning to classify images using supervised learning or learning to master the game of Go by reinforcement learning, machine reasoning can help us to integrate intent into the process.
Benefits of machine reasoning
Reliable even with limited or never seen before data
Machine reasoning can identify a brand-new piece of equipment, such as an antenna, by recognizing the similarities with older antennas. It can even make recommendations for implementing unique 5G use cases.
Explainable and auditable "thought process"
Humans wishing to audit complex network recommendations are able to follow the process taken by machine reasoning, including which information was accessed and the reasons behind any decisions made. This allows us to trust machine reasoning models.
Resolves conflicting recommendations
Communications service providers may utilize multiple machine learning models in different parts of the network, which all produce different recommendations to the same query. Machine reasoning can review these and use other data sources, as well as "common sense”, to find the best option.
Machine reasoning: How it works
Machine reasoning is built from a knowledge base and reasoner (or reasoning engine) which employs logical techniques such as deduction and induction to generate conclusions.
The knowledge base is fed by various data sources: learnings from previous solutions; codified telecom domain expertise; product documentation; external data sources; and importantly – insights from all the machine learning agents in the network.
The reasoner, using the knowledge base, looks at the various predictions, applies logical rules to the knowledge base and makes sure that business intent is followed while respecting all service level agreements and recommending a path towards the goal.
